Spark源码学习笔记5-RpcEnv(Rpc抽象层)

继4-SparkEnv 之后，我们再来详细研究下SparkEnv中出现的一个核心RpcEnv。
首先看下RpcEnv.scala源文件：
主要有RpcEnv伴生对象和伴生类，伴生类为抽象类，主要是一些Rpc框架的接口函数的申明。伴生对象提供静态的create接口，调用具体的RpcEnv工厂（NettyRpcEnvFactory）创建RpcEnv。目前Spark(我使用的是2.1.0版本)的Rpc框架的实现使用的是Netty，已经没有Akka的实现了。

package org.apache.spark.rpc
......
/**
 * A RpcEnv implementation must have a [[RpcEnvFactory]] implementation with an empty constructor
 * so that it can be created via Reflection.
 */
private[spark] object RpcEnv {

  def create(
      name: String,
      host: String,
      port: Int,
      conf: SparkConf,
      securityManager: SecurityManager,
      clientMode: Boolean = false): RpcEnv = {
    create(name, host, host, port, conf, securityManager, clientMode)
  }

  def create(
      name: String,
      bindAddress: String,
      advertiseAddress: String,
      port: Int,
      conf: SparkConf,
      securityManager: SecurityManager,
      clientMode: Boolean): RpcEnv = {
    val config = RpcEnvConfig(conf, name, bindAddress, advertiseAddress, port, securityManager,
      clientMode)
    new NettyRpcEnvFactory().create(config)
  }
}

/**
 * An RPC environment. [[RpcEndpoint]]s need to register itself with a name to [[RpcEnv]] to
 * receives messages. Then [[RpcEnv]] will process messages sent from [[RpcEndpointRef]] or remote
 * nodes, and deliver them to corresponding [[RpcEndpoint]]s. For uncaught exceptions caught by
 * [[RpcEnv]], [[RpcEnv]] will use [[RpcCallContext.sendFailure]] to send exceptions back to the
 * sender, or logging them if no such sender or `NotSerializableException`.
 *
 * [[RpcEnv]] also provides some methods to retrieve [[RpcEndpointRef]]s given name or uri.
 */
private[spark] abstract class RpcEnv(conf: SparkConf) {

  private[spark] val defaultLookupTimeout = RpcUtils.lookupRpcTimeout(conf)

  /**
   * Return RpcEndpointRef of the registered [[RpcEndpoint]]. Will be used to implement
   * [[RpcEndpoint.self]]. Return `null` if the corresponding [[RpcEndpointRef]] does not exist.
   */
  private[rpc] def endpointRef(endpoint: RpcEndpoint): RpcEndpointRef

  /**
   * Return the address that [[RpcEnv]] is listening to.
   */
  def address: RpcAddress

  /**
   * Register a [[RpcEndpoint]] with a name and return its [[RpcEndpointRef]]. [[RpcEnv]] does not
   * guarantee thread-safety.
   */
  def setupEndpoint(name: String, endpoint: RpcEndpoint): RpcEndpointRef

  /**
   * Retrieve the [[RpcEndpointRef]] represented by `uri` asynchronously.
   */
  def asyncSetupEndpointRefByURI(uri: String): Future[RpcEndpointRef]

  /**
   * Retrieve the [[RpcEndpointRef]] represented by `uri`. This is a blocking action.
   */
  def setupEndpointRefByURI(uri: String): RpcEndpointRef = {
    defaultLookupTimeout.awaitResult(asyncSetupEndpointRefByURI(uri))
  }

  /**
   * Retrieve the [[RpcEndpointRef]] represented by `address` and `endpointName`.
   * This is a blocking action.
   */
  def setupEndpointRef(address: RpcAddress, endpointName: String): RpcEndpointRef = {
    setupEndpointRefByURI(RpcEndpointAddress(address, endpointName).toString)
  }

  /**
   * Stop [[RpcEndpoint]] specified by `endpoint`.
   */
  def stop(endpoint: RpcEndpointRef): Unit

  /**
   * Shutdown this [[RpcEnv]] asynchronously. If need to make sure [[RpcEnv]] exits successfully,
   * call [[awaitTermination()]] straight after [[shutdown()]].
   */
  def shutdown(): Unit

  /**
   * Wait until [[RpcEnv]] exits.
   *
   * TODO do we need a timeout parameter?
   */
  def awaitTermination(): Unit

  /**
   * [[RpcEndpointRef]] cannot be deserialized without [[RpcEnv]]. So when deserializing any object
   * that contains [[RpcEndpointRef]]s, the deserialization codes should be wrapped by this method.
   */
  def deserialize[T](deserializationAction: () => T): T

  /**
   * Return the instance of the file server used to serve files. This may be `null` if the
   * RpcEnv is not operating in server mode.
   */
  def fileServer: RpcEnvFileServer

  /**
   * Open a channel to download a file from the given URI. If the URIs returned by the
   * RpcEnvFileServer use the "spark" scheme, this method will be called by the Utils class to
   * retrieve the files.
   *
   * @param uri URI with location of the file.
   */
  def openChannel(uri: String): ReadableByteChannel
}

在RpcEnv.scala源文件中除了RpcEnv的伴生对象和伴生类外，还有一个样例类RpcEnvConfig和一个trait RpcEnvFileServer。RpcEnvConfig的实例作为参数传递给具体的RpcEnvFactory的create函数创建RpcEnv实例：

private[spark] case class RpcEnvConfig(
    conf: SparkConf,
    name: String,
    bindAddress: String,
    advertiseAddress: String,
    port: Int,
    securityManager: SecurityManager,
    clientMode: Boolean)

/**
 * A server used by the RpcEnv to server files to other processes owned by the application.
 *
 * The file server can return URIs handled by common libraries (such as "http" or "hdfs"), or
 * it can return "spark" URIs which will be handled by `RpcEnv#fetchFile`.
 */
private[spark] trait RpcEnvFileServer {

  /**
   * Adds a file to be served by this RpcEnv. This is used to serve files from the driver
   * to executors when they're stored on the driver's local file system.
   ......
   */
  def addFile(file: File): String

  /**
   * Adds a jar to be served by this RpcEnv. 
   ......
   */
  def addJar(file: File): String

  /**
   * Adds a local directory to be served via this file server.
   * ......
   */
  def addDirectory(baseUri: String, path: File): String

  /** Validates and normalizes the base URI for directories. */
  protected def validateDirectoryUri(baseUri: String): String = {
    val fixedBaseUri = "/" + baseUri.stripPrefix("/").stripSuffix("/")
    require(fixedBaseUri != "/files" && fixedBaseUri != "/jars",
      "Directory URI cannot be /files nor /jars.")
    fixedBaseUri
  }

}

RpcEnv.scala文件中已涉及类NettyRpcEnvFactory, RpcEndpoint, RpcEndpointRef。除了NettyRpcEnvFactory是Rpc框架实现层的类外，RpcEndpoint，RpcEndpointRef以及RpcEnv都是Rpc框架抽象层的类，这一节我们主要了解抽象层的类。

RpcEndpoint所在文件为RpcEndpoint.scala，该文件包含trait RpcEnvFactory的定义：

/**
 * A factory class to create the [[RpcEnv]]. It must have an empty constructor so that it can be
 * created using Reflection.
 */
private[spark] trait RpcEnvFactory {

  def create(config: RpcEnvConfig): RpcEnv
}

然后主要是trait RpcEndpoint的定义，RpcEndpoint是Rpc通讯的端点，其trait包含启动/接受消息/停止/连接状态变化的处理等函数的default实现：

/**
 * An end point for the RPC that defines what functions to trigger given a message.
 *
 * It is guaranteed that `onStart`, `receive` and `onStop` will be called in sequence.
 *
 * The life-cycle of an endpoint is:
 *
 * constructor -> onStart -> receive* -> onStop
 *
 * Note: `receive` can be called concurrently. If you want `receive` to be thread-safe, please use
 * [[ThreadSafeRpcEndpoint]]
 *
 * If any error is thrown from one of [[RpcEndpoint]] methods except `onError`, `onError` will be
 * invoked with the cause. If `onError` throws an error, [[RpcEnv]] will ignore it.
 */
private[spark] trait RpcEndpoint {

  /**
   * The [[RpcEnv]] that this [[RpcEndpoint]] is registered to.
   */
  val rpcEnv: RpcEnv

  /**
   * The [[RpcEndpointRef]] of this [[RpcEndpoint]]. `self` will become valid when `onStart` is
   * called. And `self` will become `null` when `onStop` is called.
   *
   * Note: Because before `onStart`, [[RpcEndpoint]] has not yet been registered and there is not
   * valid [[RpcEndpointRef]] for it. So don't call `self` before `onStart` is called.
   */
  final def self: RpcEndpointRef = {
    require(rpcEnv != null, "rpcEnv has not been initialized")
    rpcEnv.endpointRef(this)
  }

  /**
   * Process messages from [[RpcEndpointRef.send]] or [[RpcCallContext.reply)]]. If receiving a
   * unmatched message, [[SparkException]] will be thrown and sent to `onError`.
   */
  def receive: PartialFunction[Any, Unit] = {
    case _ => throw new SparkException(self + " does not implement 'receive'")
  }

  /**
   * Process messages from [[RpcEndpointRef.ask]]. If receiving a unmatched message,
   * [[SparkException]] will be thrown and sent to `onError`.
   */
  def receiveAndReply(context: RpcCallContext): PartialFunction[Any, Unit] = {
    case _ => context.sendFailure(new SparkException(self + " won't reply anything"))
  }

  /**
   * Invoked when any exception is thrown during handling messages.
   */
  def onError(cause: Throwable): Unit = {
    // By default, throw e and let RpcEnv handle it
    throw cause
  }

  /**
   * Invoked when `remoteAddress` is connected to the current node.
   */
  def onConnected(remoteAddress: RpcAddress): Unit = {
    // By default, do nothing.
  }

  /**
   * Invoked when `remoteAddress` is lost.
   */
  def onDisconnected(remoteAddress: RpcAddress): Unit = {
    // By default, do nothing.
  }

  /**
   * Invoked when some network error happens in the connection between the current node and
   * `remoteAddress`.
   */
  def onNetworkError(cause: Throwable, remoteAddress: RpcAddress): Unit = {
    // By default, do nothing.
  }

  /**
   * Invoked before [[RpcEndpoint]] starts to handle any message.
   */
  def onStart(): Unit = {
    // By default, do nothing.
  }

  /**
   * Invoked when [[RpcEndpoint]] is stopping. `self` will be `null` in this method and you cannot
   * use it to send or ask messages.
   */
  def onStop(): Unit = {
    // By default, do nothing.
  }

  /**
   * A convenient method to stop [[RpcEndpoint]].
   */
  final def stop(): Unit = {
    val _self = self
    if (_self != null) {
      rpcEnv.stop(_self)
    }
  }
}

/**
 * A trait that requires RpcEnv thread-safely sending messages to it.
 *
 ......
 */
private[spark] trait ThreadSafeRpcEndpoint extends RpcEndpoint

RpcEndpointRef所在文件为RpcEndpointRef.scala, 该文件仅包含RpcEndpointRef这个类，该类是一个抽象类，它是对一个远程RpcEndpoint的引用，通过一些接口与远端RpcEndpoint通讯。它与RpcEndpoint一样也是Rpc框架抽象层的类，定义如下：

package org.apache.spark.rpc
......
/**
 * A reference for a remote [[RpcEndpoint]]. [[RpcEndpointRef]] is thread-safe.
 */
private[spark] abstract class RpcEndpointRef(conf: SparkConf)
  extends Serializable with Logging {

  private[this] val maxRetries = RpcUtils.numRetries(conf)
  private[this] val retryWaitMs = RpcUtils.retryWaitMs(conf)
  private[this] val defaultAskTimeout = RpcUtils.askRpcTimeout(conf)

  /**
   * return the address for the [[RpcEndpointRef]]
   */
  def address: RpcAddress

  def name: String

  /**
   * Sends a one-way asynchronous message. Fire-and-forget semantics.
   */
  def send(message: Any): Unit

  /**
   * Send a message to the corresponding [[RpcEndpoint.receiveAndReply)]] and return a [[Future]] to
   * receive the reply within the specified timeout.
   *
   * This method only sends the message once and never retries.
   */
  def ask[T: ClassTag](message: Any, timeout: RpcTimeout): Future[T]

  /**
   * Send a message to the corresponding [[RpcEndpoint.receiveAndReply)]] and return a [[Future]] to
   * receive the reply within a default timeout.
   *
   * This method only sends the message once and never retries.
   */
  def ask[T: ClassTag](message: Any): Future[T] = ask(message, defaultAskTimeout)

  /**
   * Send a message to the corresponding [[RpcEndpoint]] and get its result within a default
   * timeout, or throw a SparkException if this fails even after the default number of retries.
   * The default `timeout` will be used in every trial of calling `sendWithReply`. Because this
   * method retries, the message handling in the receiver side should be idempotent.
   *
   * Note: this is a blocking action which may cost a lot of time,  so don't call it in a message
   * loop of [[RpcEndpoint]].
   *
   * @param message the message to send
   * @tparam T type of the reply message
   * @return the reply message from the corresponding [[RpcEndpoint]]
   */
  def askWithRetry[T: ClassTag](message: Any): T = askWithRetry(message, defaultAskTimeout)

  /**
   * Send a message to the corresponding [[RpcEndpoint.receive]] and get its result within a
   * specified timeout, throw a SparkException if this fails even after the specified number of
   * retries. `timeout` will be used in every trial of calling `sendWithReply`. Because this method
   * retries, the message handling in the receiver side should be idempotent.
   *
   * Note: this is a blocking action which may cost a lot of time, so don't call it in a message
   * loop of [[RpcEndpoint]].
   *
   * @param message the message to send
   * @param timeout the timeout duration
   * @tparam T type of the reply message
   * @return the reply message from the corresponding [[RpcEndpoint]]
   */
  def askWithRetry[T: ClassTag](message: Any, timeout: RpcTimeout): T = {
    // TODO: Consider removing multiple attempts
    var attempts = 0
    var lastException: Exception = null
    while (attempts < maxRetries) {
      attempts += 1
      try {
        val future = ask[T](message, timeout)
        val result = timeout.awaitResult(future)
        if (result == null) {
          throw new SparkException("RpcEndpoint returned null")
        }
        return result
      } catch {
        case ie: InterruptedException => throw ie
        case e: Exception =>
          lastException = e
          logWarning(s"Error sending message [message = $message] in $attempts attempts", e)
      }

      if (attempts < maxRetries) {
        Thread.sleep(retryWaitMs)
      }
    }

    throw new SparkException(
      s"Error sending message [message = $message]", lastException)
  }

}